What is MCP analytics, in one sentence?

MCP analytics is a website analytics surface exposed through Model Context Protocol so any MCP-compatible client (Claude Desktop, Claude Code, the VS Code Claude extension) can read traffic, attribution, and AI visibility data on the operator's behalf — read-only, workspace-scoped, audit-logged.

Which Claude clients can read WebmasterID through MCP?

Any client that speaks MCP today. The /settings/mcp page in the dashboard ships ready-to-paste configuration snippets for Claude Desktop, Claude Code (CLI), and the VS Code Claude extension. The endpoint is workspace-scoped and authenticated by a Bearer API key.

Is data shared with model providers?

No. When you ask Claude to read your data through MCP, the request goes directly from your Claude client to our MCP endpoint — we do not relay analytics into a model provider for training. The data is yours and stays workspace-scoped.

Why MCP and not a custom REST API?

Model Context Protocol is an open standard for connecting AI assistants to external data and tools. Using MCP means the same analytics surface works across many AI clients without a per-vendor adapter. It also keeps the integration auditable: every MCP call lands in the workspace audit log.

What does setup look like?

Generate a workspace-scoped API key in the dashboard at /settings/mcp, copy the client-specific snippet, restart your Claude client. The WebmasterID tools appear in the client's available-tools list. Total setup is under five minutes.

MCP analytics

Website analytics that Claude can read — through Model Context Protocol

WebmasterID exposes traffic, attribution, AI visibility, and Agent recommendations through MCP. Workspace-scoped Bearer auth, read-only surface, every read recorded in the audit log.

Start free See pricing Technical detail

Use cases

What MCP analytics is for

The use cases below are the day-to-day shape of the workflow. Each one is a prompt the operator runs alongside the dashboard, not a replacement for it.

Ask Claude about your traffic

“Which AI assistants sent traffic to /pricing this week?” One prompt, MCP fetches the relevant slice, Claude answers grounded in data.

Drill into a specific source

Walk from a workspace-level attribution rollup down to the underlying events without leaving the conversation.

Compare two date ranges

Week-over-week or month-over-month comparisons in a turn. The data is fetched through MCP; Claude summarises what changed.

Read the Agent recommendations

MCP exposes the WebmasterID Agent surface as well — recommendations, repo mappings, and the daily operator brief.

Investigate an incident

Tracker not firing? Ask Claude to read the recent events for a site. The audit log records every MCP read.

Brief a teammate

Hand the conversation to a teammate; the MCP-read data is reproducible. No screenshots, no stale dashboards.

How WebmasterID helps

Model Context Protocol is the integration layer

MCP is the open standard for connecting AI assistants to external data and tools. WebmasterID ships an MCP server; Claude (or any MCP client) connects with a workspace-scoped Bearer key and reads the allow-listed surface.

1. Generate a key. At /settings/mcp in the dashboard, generate a workspace-scoped API key.
2. Paste the snippet. Copy the client-specific snippet for Claude Desktop, Claude Code, or the VS Code Claude extension. Restart the client.
3. Ask analytics questions. The WebmasterID tools appear in the available-tools list. Claude can call them on your behalf and read the result.
4. Audit and review. Every MCP call lands in the workspace audit log. You see what was read, when, and by which API key.

Safety model. MCP gives Claude read access and prompt-preparation capability. Code changes, plan changes, and configuration changes are reviewed and deployed by the operator. There is no auto-deploy and no hidden execution.

Feature

What the MCP surface exposes

A small, deterministic surface. No fabricated tools, no fake autonomy.

Workspace + sites

Workspace metadata, the site list, and per-site summaries — enough context for Claude to ground questions correctly.

Traffic + events

Recent events filtered by date, traffic type, country, pathname, referrer, session, or visitor. The same surface that powers the Event Explorer.

AI visibility

AI crawler timeline and AI assistant referrals. Useful for AI-search-era investigations.

Attribution

Directional attribution rollups with source-confidence labels. Claude can read and summarise; the operator decides.

Agent recommendations

WebmasterID Agent recommendations and the daily operator brief. Helpful for AI-assisted prioritisation.

Audit log

Every read is recorded. Operators see what was fetched, when, and by which Bearer key.

Privacy & trust

The integration is auditable by construction

Operators evaluating MCP integrations rightly worry about silent data exfiltration. WebmasterID's MCP surface is built to remove that worry.

Workspace-scoped
Bearer keys are bound to a single workspace. Cross-workspace reads are impossible.
Read-only
No write tools today. The MCP server cannot mutate data, cannot deploy, cannot charge or refund.
Audit-logged
Each read appends to the workspace audit log. Operators see what Claude (or any other MCP client) requested.
No model training
We do not train shared models on workspace data. The data is yours; the integration is read-only on your behalf.

Questions about the MCP integration? info@helperg.com.

FAQ

MCP analytics, answered

What is MCP analytics, in one sentence?: MCP analytics is a website analytics surface exposed through Model Context Protocol so any MCP-compatible client (Claude Desktop, Claude Code, the VS Code Claude extension) can read traffic, attribution, and AI visibility data on the operator's behalf — read-only, workspace-scoped, audit-logged.
Which Claude clients can read WebmasterID through MCP?: Any client that speaks MCP today. The /settings/mcp page in the dashboard ships ready-to-paste configuration snippets for Claude Desktop, Claude Code (CLI), and the VS Code Claude extension. The endpoint is workspace-scoped and authenticated by a Bearer API key.
Is the MCP surface read-only?: Yes. The MCP server exposes read tools today — there are no write tools. Code changes, plan changes, and configuration changes are reviewed and deployed by the operator. There is no auto-deploy, no auto-merge, no hidden execution.
Is data shared with model providers?: No. When you ask Claude to read your data through MCP, the request goes directly from your Claude client to our MCP endpoint — we do not relay analytics into a model provider for training. The data is yours and stays workspace-scoped.
Why MCP and not a custom REST API?: Model Context Protocol is an open standard for connecting AI assistants to external data and tools. Using MCP means the same analytics surface works across many AI clients without a per-vendor adapter. It also keeps the integration auditable: every MCP call lands in the workspace audit log.
What does setup look like?: Generate a workspace-scoped API key in the dashboard at /settings/mcp, copy the client-specific snippet, restart your Claude client. The WebmasterID tools appear in the client's available-tools list. Total setup is under five minutes.

Related cluster pages

Same site, deeper coverage of the pieces MCP analytics depends on.

Claude analytics integration

The operator's view of Claude Desktop, Claude Code, and the VS Code Claude extension reading WebmasterID.

AI-assisted analytics

The product story around the AI-assisted workflow shape.

AI visibility analytics

The data Claude reads first — AI crawlers and AI-assistant referrals.

Event Explorer

The investigation surface that the MCP server mirrors for AI clients.

Multi-site analytics

MCP is workspace-scoped — one Bearer key reads the whole portfolio.

Docs

Configure the MCP integration end-to-end.